Bug#338459: On an HP nx6125, ACPI thermal events are not processed until one does an acpi -t
I can now confirm that this bug exists also in the latest (testing) Debian kernel: linux-image-2.6.14-1-amd64-k8. I booted with
boot options (i) disable_timer_pin_1 and (ii) disable_timer_pin_1 together with ec_burst=1. In both instances my fans only worked when I issued an acpi -t.
It is perhaps worthwhile mentioning that the behaviour is not entirely symmetric in that the fans usually (sometimes) switch off when the CPU temp decreases below a thermal trip point. However, when a trip point is exceeded they almost always fail to turn on. I suspect this has to do with CPU load . You exceed a thermal trip point when you "work" your CPU. Your CPU cools down when the CPU load diminishes from a previously high workload state. (This seems to confirm my comments in the original report regarding the occurence of this bug under moderate to high CPU load.)
I answer some of your questions below, Frederik...
>>> Frederik Schueler <fs@lowpingbastards.de> 11/10/05 9:53 PM >>>
>>are you sure the fan module is loaded?
I've compiled ACPI support for fans into the vanilla kernels :
--------exerpt from my 2.6.14.1 .config-----------------
CONFIG_ACPI=y
CONFIG_ACPI_SLEEP=y
CONFIG_ACPI_SLEEP_PROC_FS=y
# CONFIG_ACPI_SLEEP_PROC_SLEEP is not set
CONFIG_ACPI_AC=y
CONFIG_ACPI_BATTERY=y
CONFIG_ACPI_BUTTON=y
CONFIG_ACPI_VIDEO=y
CONFIG_ACPI_HOTKEY=m
CONFIG_ACPI_FAN=y
CONFIG_ACPI_PROCESSOR=y
CONFIG_ACPI_THERMAL=y
CONFIG_ACPI_ASUS=m
CONFIG_ACPI_IBM=m
CONFIG_ACPI_TOSHIBA=m
CONFIG_ACPI_BLACKLIST_YEAR=0
# CONFIG_ACPI_DEBUG is not set
CONFIG_ACPI_EC=y
CONFIG_ACPI_POWER=y
CONFIG_ACPI_SYSTEM=y
CONFIG_ACPI_CONTAINER=m
--------------------end excerpt--------------------
and when I use the standard Debian kernels, acpid loads the necessary modules, as far as I can tell.
>> Any other advice you can offer? Can I provide you with any other information?
>you can give an acpi-patched 2.6.13 a try - get the patch from
>http://ftp.kernel.org/pub/linux/kernel/people/lenb/acpi/patches/release/
Thanks. I will give anything a try at this point!!
>most of the patch was merged in 2.6.14, but it should be worth a >try.
>there is no patch for 2.6.14 yet.
Thanks for the info.
>The only workaround which comes to mind is booting with >acpi=off until
>a fix shows up.
If memory serves me correcly, this causes a kernel panic on boot with the HP nx6125, so is not an option for me.
>In the meantime, you should seek help on the acpi-devel
>mailinglist.
>http://acpi.sourceforge.net/mailinglists.html
I will certainly try there. Many thanks for the pointers and help.
Best wishes
Richard
(P.S. thanks for the information about disable_timer_pin_1. I wasn't aware of this option and it works like a charm. Generates fewer errors than no_timer_check.)
--
ENOSIG
--------------------------------------------------------------------
Please find our disclaimer at http://www.ukzn.ac.za/disclaimer
--------------------------------------------------------------------
<<<<gwavasig>>>>
Reply to: